Add optimizer control settings for classical LMM and GLMM

[Copilot]

This PR adds optimizer control settings to the classical Linear Mixed Models (LMM) and Generalized Linear Mixed Models (GLMM) modules, allowing users to adjust optimizer parameters when models fail to converge.

Changes Made

UI Enhancements

• Added "Advanced" section with optimizer settings as a separate QML component
• Optimizer-specific controls that appear based on selected optimizer method:
  • Nelder-Mead: maxfun, FtolAbs, FtolRel, XtolRel
  • bobyqa: npt, rhobeg, rhoend, maxfun
  • nlminb: tol, relTol, maxit
  • Default/BFGS: Maximum iterations, Convergence tolerance
  • Check convergence: Toggle for convergence warnings/errors (available for all optimizers)

Code Organization

• Created a reusable Advanced.qml component in the common folder containing optimizer settings
• This component is added to the bottom of both LMM and GLMM QML files for better modularity
• Removed optimizer settings from the main Options section to improve code separation

Backend Implementation

• Updated .mmCreateOptimizerControl() function to handle optimizer-specific parameters according to each optimizer's documentation
• Updated all fitting functions to use optimizer controls:
  • .mixedInterceptML() for intercept-only models
  • afex::mixed() calls for full models
• Controls are passed through to underlying lmer/glmer functions via optCtrl parameter

Error Message Improvements

• Updated convergence failure message to direct users to optimizer settings instead of stating they're not supported
• Changed from: "(JASP currently does not support modifying the optimizer settings.)"
• To: "Try adjusting the optimizer settings in the Advanced Options section."

Testing

• Added unit tests for both LMM and GLMM with custom optimizer settings
• Tests validate that new options are processed without errors

Scope

This enhancement is only available for classical models (LMM and GLMM). Bayesian models continue to use their existing MCMC control settings and are unaffected by these changes.

Motivation

Complex mixed models often encounter convergence issues that can be resolved by adjusting optimizer settings. This was previously impossible in JASP, forcing users to abandon analyses or switch to other tools. These new controls provide the flexibility needed to handle challenging models while maintaining JASP's user-friendly interface.

Fixes #252.

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.

[FBartos]

@copilot put the QML interface into a new common separate qml section called Advanced thats added the bottom of the LMM and GLMM's qml

[Copilot]

@copilot put the QML interface into a new common separate qml section called Advanced thats added the bottom of the LMM and GLMM's qml

Done! I've extracted the optimizer settings into a new common QML component called Advanced.qml in the common folder. This component is now added to the bottom of both LMM and GLMM QML files instead of being part of the Options section. The changes are committed in 7256221.

[FBartos]

so, this turned to be much more complicated than expected